{"id":{"repo_id":"uic","oai_identifier":"oai:figshare.com:article/32993888"},"canonical_url":"https://search.dev.ndltd.org/etd/uic/oai:figshare.com:article/32993888","repository":{"repo_id":"uic","name":"University of Illinois - Chicago","base_url":"https://api.figshare.com/v2/oai"},"display":{"title":"Applications of Optimization and Machine Learning in Assortment Optimization and Revenue Management","abstract":"Firms in modern product markets must often make assortment and pricing decisions in environments where demand structure is complex, high-dimensional, or only partially observable. Traditional approaches typically rely on parametric demand models and rich transactional data, which may not always be available. This dissertation develops new methodological approaches that combine optimization, robust decision frameworks, and machine learning to support decision-making under such conditions. The first chapter studies assortment optimization under the random utility maximization model using the sample average approximation. To improve the practical viability of this approach, the chapter introduces a stronger mixed-integer programming formulation and an accelerated Benders decomposition method that substantially reduces computation times in large-scale problems. The second chapter develops a sales-aware product embedding model trained using publicly observable Amazon review dynamics. By focusing on relative performance within competitive product sets, the model learns product representations that better capture substitution patterns and improve downstream demand prediction tasks. The third chapter proposes a nonparametric framework for multi-product pricing that avoids estimating a parametric demand model. Pricing decisions are formulated as a robust optimization problem that maximizes worst-case revenue over demand functions consistent with observed data and basic rationality assumptions. Together, these contributions expand the methodological tools available for analyzing and solving decision problems in product markets where demand information is incomplete or difficult to model parametrically.","abstract_html":"Firms in modern product markets must often make assortment and pricing decisions in environments where demand structure is complex, high-dimensional, or only partially observable. Traditional approaches typically rely on parametric demand models and rich transactional data, which may not always be available. This dissertation develops new methodological approaches that combine optimization, robust decision frameworks, and machine learning to support decision-making under such conditions. The first chapter studies assortment optimization under the random utility maximization model using the sample average approximation. To improve the practical viability of this approach, the chapter introduces a stronger mixed-integer programming formulation and an accelerated Benders decomposition method that substantially reduces computation times in large-scale problems. The second chapter develops a sales-aware product embedding model trained using publicly observable Amazon review dynamics. By focusing on relative performance within competitive product sets, the model learns product representations that better capture substitution patterns and improve downstream demand prediction tasks. The third chapter proposes a nonparametric framework for multi-product pricing that avoids estimating a parametric demand model. Pricing decisions are formulated as a robust optimization problem that maximizes worst-case revenue over demand functions consistent with observed data and basic rationality assumptions. Together, these contributions expand the methodological tools available for analyzing and solving decision problems in product markets where demand information is incomplete or difficult to model parametrically.","abstract_has_math":false,"creators":["Hassaan Khalid (24399455)"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-05-01T00:00:00Z","date_published":"2026-05-01T00:00:00Z","updated_at":"2026-07-27T21:33:40Z","subjects":["Operations Research","Machine Learning","Optimization","Revenue Management","Data Science"],"languages":[],"rights":["In Copyright"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://doi.org/10.25417/uic.32993888.v1","outbound_label":"DOI","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:creator","label":"Author","values":["Hassaan Khalid (24399455)"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026-05-01T00:00:00Z"]},{"key":"dc:relation","label":"Dc Relation","values":["https://figshare.com/articles/thesis/Applications_of_Optimization_and_Machine_Learning_in_Assortment_Optimization_and_Revenue_Management/32993888"]},{"key":"dc:type","label":"Dc Type","values":["Text","Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Operations Research","Machine Learning","Optimization","Revenue Management","Data Science"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:rights","label":"Dc Rights","values":["In Copyright"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["10.25417/uic.32993888.v1"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Firms in modern product markets must often make assortment and pricing decisions in environments where demand structure is complex, high-dimensional, or only partially observable. Traditional approaches typically rely on parametric demand models and rich transactional data, which may not always be available. This dissertation develops new methodological approaches that combine optimization, robust decision frameworks, and machine learning to support decision-making under such conditions. The first chapter studies assortment optimization under the random utility maximization model using the sample average approximation. To improve the practical viability of this approach, the chapter introduces a stronger mixed-integer programming formulation and an accelerated Benders decomposition method that substantially reduces computation times in large-scale problems. The second chapter develops a sales-aware product embedding model trained using publicly observable Amazon review dynamics. By focusing on relative performance within competitive product sets, the model learns product representations that better capture substitution patterns and improve downstream demand prediction tasks. The third chapter proposes a nonparametric framework for multi-product pricing that avoids estimating a parametric demand model. Pricing decisions are formulated as a robust optimization problem that maximizes worst-case revenue over demand functions consistent with observed data and basic rationality assumptions. Together, these contributions expand the methodological tools available for analyzing and solving decision problems in product markets where demand information is incomplete or difficult to model parametrically."]},{"key":"dc:title","label":"Title","values":["Applications of Optimization and Machine Learning in Assortment Optimization and Revenue Management"]}]}],"canonical_facts":{"dc:creator":["Hassaan Khalid (24399455)"],"dc:date":["2026-05-01T00:00:00Z"],"dc:description":["Firms in modern product markets must often make assortment and pricing decisions in environments where demand structure is complex, high-dimensional, or only partially observable. Traditional approaches typically rely on parametric demand models and rich transactional data, which may not always be available. This dissertation develops new methodological approaches that combine optimization, robust decision frameworks, and machine learning to support decision-making under such conditions. The first chapter studies assortment optimization under the random utility maximization model using the sample average approximation. To improve the practical viability of this approach, the chapter introduces a stronger mixed-integer programming formulation and an accelerated Benders decomposition method that substantially reduces computation times in large-scale problems. The second chapter develops a sales-aware product embedding model trained using publicly observable Amazon review dynamics. By focusing on relative performance within competitive product sets, the model learns product representations that better capture substitution patterns and improve downstream demand prediction tasks. The third chapter proposes a nonparametric framework for multi-product pricing that avoids estimating a parametric demand model. Pricing decisions are formulated as a robust optimization problem that maximizes worst-case revenue over demand functions consistent with observed data and basic rationality assumptions. Together, these contributions expand the methodological tools available for analyzing and solving decision problems in product markets where demand information is incomplete or difficult to model parametrically."],"dc:identifier":["10.25417/uic.32993888.v1"],"dc:relation":["https://figshare.com/articles/thesis/Applications_of_Optimization_and_Machine_Learning_in_Assortment_Optimization_and_Revenue_Management/32993888"],"dc:rights":["In Copyright"],"dc:subject":["Operations Research","Machine Learning","Optimization","Revenue Management","Data Science"],"dc:title":["Applications of Optimization and Machine Learning in Assortment Optimization and Revenue Management"],"dc:type":["Text","Thesis"]},"updated_at":"2026-07-27T21:33:40Z"}